Human cytomegalovirus (HCMV) is a ubiquitous human pathogen. While the virus has generally low pathogen potential in normal individuals, it causes variety of debilitating and life-threatening illnesses (including ulcerative colitis, pneumonia, and chorioretinitis) in newborns and in immunocompromised individuals. Because of the profound immunosuppression in HIV infection, HCMV induced disease can be especially severe in AIDS patients. At the present time there are two agents which are licensed for the treatment of HCMV infection, ganciclovir and Foscarnet. Although both drugs have found to be clinically useful, the general utility of these drugs is limited by the need for intravenous administration and by low therapeutic/toxicity ratios. The increasing morbidity and mortality caused by both primary and recurring HCMV infections in AIDS has created a desperate need to anti-HCMV agents with improved efficacy. We propose to use a combination of x-ray crystallographic and computational modelling studies of the proteins which are the targets of current anti-CMV agents to design compounds with improved bioavailability and therapeutic/toxicity ratios. Two general strategies will be used: 1) To use structures of complexes of the protein targets with analogues of existing drugs to design modifications of the current drugs with improved binding and/or specificity. 2)To use the structure of the target molecule to identify new lead compounds. Our initial studies will focus on the product of the UL97 gene of HCMV. This protein, which has significant sequence homologies with protein kinases, has recently been shown to be capable of phosphorylating ganciclovir and is essential for its antiviral activity. The specific aims of this portion of the project will be to: 1) To build a preliminary model for the catalytic domain of the UL97 gene product based on sequence homologies with the protein kinases and on the known structure of the catalytic domain of the cAMP-dependent protein kinase. 2) To purify the UL97 gene product and/or active fragments (to be supplied by Dr. Donald Cohen). 3) To crystallize the UL97 gene product and/or active fragments. 4) To solve and refine the structures of these proteins free and in complex with MgATP, ganciclovir, or other substrate analogues. 5) To use the structure of the protein and its complexes to design agents with improved anti-HCMV activity. This last goal will be carried out in collaboration with Burroughs Wellcome. As the project proceeds we will expand the structural studies to include the HCMV DNA polymerase, its active fragments, and accessory proteins (e.g. ICP36) Genetic and biochemical characterization of these proteins, and their overexpression for crystallographic studies is the subject of project 0002 in this program.